FRET Benchmark

This directory contains all the scripts to run the benchmark of the FRET_R Bayesian restraint, and the reference results.

• DATA:

  • GMM: Gaussian Mixture Models needed by model for linker flexibility
  • PDBs: Pdb with complex structure, one per chain
  • REFERENCE_RESULTS: reference results of benchmark
  • SCRIPTS: scripts to run and analyze the benchmark
    • ANALYSIS: scripts to analyze the results
    • BENCHMARK_LINKER: scripts to run the benchmark with model for linker flexibility
    • BENCHMARK_NOLINKER: scripts to run the benchmark without model for linker flexibility

• RESULTS_LINKER : directory where the results of benchmark with model for linker flexibility will be stored

• RESULTS_NOLINKER : directory where the results of benchmark without model for linker flexibility will be stored

0. This is the ROOT directory.

1. PRELIMINARY. Compile IMP (any recent version should work - see below for the last known good IMP version). IMP git version develop-c47408c was used for the publication.

2. EXECUTION. Go to DATA/SCRIPTS/BENCHMARK_LINKER or DATA/SCRIPTS/BENCHMARK_NOLINKER where you will find the scripts to run the benchmark with or without using a model for linker flexibility, respectively. Then:

   • To run a specific test, run:

     ./do_test.sh ID

     where ID is an integer from 1 to 1280.

   • To run all tests, just use a bash loop:

     for((ID=1;ID<=1280;ID++)); do ./do_test.sh $ID; done

   • A sample SGE script to run all the tests on a cluster is also provided, as job.sh.

3. ANALYSIS. From the ROOT directory, do:

   • DATA/SCRIPTS/ANALYSIS/do_analysis.py RESULTS_LINKER linker.dat
   • DATA/SCRIPTS/ANALYSIS/do_analysis.py RESULTS_NOLINKER nolinker.dat
   • DATA/SCRIPTS/ANALYSIS/do_analysis_per_complex.py RESULTS_LINKER linker_per_complex.dat
   • DATA/SCRIPTS/ANALYSIS/do_analysis_per_complex.py RESULTS_NOLINKER nolinker_per_complex.dat

   Statistics will be written to the .dat files, which should be compared to the results in DATA/REFERENCE_RESULTS and to those published in Bonomi et al. "Determining protein complex structures based on a Bayesian model of in vivo FRET data", as follows.

   • linker.dat -> Table S2
   • nolinker.dat -> Table S3

Author(s): Max Bonomi

Version: 1.0

License: LGPL. This library is free software; you can redistribute it and/or modify it under the terms of the GNU Lesser General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version.

Last known good IMP version:

Testable: Yes.

Parallelizeable: Yes

Publications:

• M. Bonomi, E.G. Muller, R. Pellarin, S.J. Kim, D. Russel, R. Ramsden, B.A. Sundin, T.A. Davis, A. Sali. Determining protein complex structures based on a Bayesian model of in vivo FRET data. Mol Cell Proteomics 13, 2812-2823, 2014.